Communication machine room three-dimensional model instantiation rendering optimization method and system

Abstract

The invention relates to a communication machine room three-dimensional model instantiation rendering optimization method and system, and the method comprises the steps: 1, inputting rendering parameter data into a machine learning platform for period prediction, and obtaining an optimal rendering period; 2, according to the optimal rendering period, dynamically distributing rendering resources and dynamically selecting a rendering method to carry out actual rendering, and obtaining actual rendering data; inputting the actual rendering data into the machine learning platform for learning parameter optimization to obtain a learning parameter optimized machine learning platform; 3, the machine learning platform with the optimized learning parameters performs period prediction according to the current rendering parameter data of the three-dimensional model of the communication machine room to obtain a next optimal rendering period; and 4, repeating the step 2 and the step 3 according to the next optimal rendering period until the actual rendering of the three-dimensional model of the communication machine room is finished. By implementing the method provided by the invention, the rendering efficiency and quality of the three-dimensional model of the communication machine room can be effectively improved.

Description

Technical Field

[0001] The invention relates to the field of computer graphics, and in particular to a communication room three-dimensional model instantiation rendering optimization method. Background Art

[0002] In the field of computer graphics, rendering of 3D models is an important step in generating images. Traditional rendering methods often require a lot of computing resources and time when processing complex scenes, which is particularly problematic in real-time rendering applications. For example, in scenarios such as video games, virtual reality, and augmented reality, high-quality 3D images need to be rendered quickly to provide a smooth user experience. However, existing rendering technologies are prone to performance bottlenecks when processing a large number of instantiated models, resulting in reduced rendering speed and image quality.

[0003] Therefore, a variety of rendering optimization methods have emerged in the prior art. For example, the patent document with application number CN202411110468.6 provides a 3D rendering method based on multi-time sequence scenes, including: collecting videos or pictures of the scene to be rendered, preprocessing the videos or pictures, using the SfM algorithm, obtaining the point cloud file and image pose information of the scene to be rendered and generating a data packet; inputting the data packet into a 3D Gaussian rendering algorithm for multi-time sequence scenes, and outputting real scene renderings and point cloud files from various perspectives; wherein the 3D Gaussian rendering algorithm specifically includes the following steps: S1, constructing Optimizer: The optimizer includes the spatial position information, spherical harmonic coefficients and transparency information of each point in the point cloud; S2, encoding: for existing transient objects, based on structured information and light intensity, transient encoding of point clouds and images of scenes to be rendered is performed; S3, optimizer update: all three-dimensional Gaussian points are generated according to the internal and external orientation elements of the input image to generate a simulation view from that perspective, combined with transient encoding, the transient part is removed, the loss function is calculated, and the optimizer is updated using gradient descent; S4, rendering: using the various parameters of the optimizer, backpropagation is performed to update the point cloud information, and finally the rendered model file is obtained.

[0004] However, although the above rendering method improves rendering efficiency to a certain extent, it still has some limitations. For example, when processing large-scale scenes, the method is not flexible enough in handling the rendering method of objects in the scene to be rendered, and cannot adjust the rendering resources and strategies according to the characteristics of the objects, resulting in difficulty in achieving the optimal balance between rendering effect and efficiency. Summary of the invention

[0005] In order to overcome the shortcomings of the prior art, the present application provides, on the one hand, a method for optimizing instantiation rendering of a three-dimensional model of a communication room, comprising: S1. Inputting the rendering parameter data of the three-dimensional model of the communication room into a machine learning platform for cycle prediction to obtain an optimal rendering cycle; the machine learning platform is constructed based on historical rendering data, and the historical rendering data includes historical rendering parameter data; S2. Dynamically allocate rendering resources and dynamically select a rendering method to perform actual rendering according to the optimal rendering cycle to obtain actual rendering data; the rendering method includes at least one instantiated rendering and at least one non-instantiated rendering; S3, inputting the actual rendering data into the machine learning platform to optimize the learning parameters, and obtaining a machine learning platform with optimized learning parameters; the machine learning platform with optimized learning parameters performs cycle prediction based on the current rendering parameter data of the three-dimensional model of the communication room, and obtains the next optimal rendering cycle; S4. Repeat steps S2 and S3 according to the next optimal rendering cycle until the actual rendering of the communication room three-dimensional model is completed.

[0006] Furthermore, the rendering parameter data of the three-dimensional model of the communication room includes: geometric parameters, material properties, lighting conditions and viewing angle information of the objects to be rendered in the communication room, the geometric parameters include size, shape and position information of the objects to be rendered; the material properties include reflectivity, transparency and texture information of the object surface; the lighting conditions include intensity, direction and color information of the light source; the viewing angle information includes the position and direction of the observer.

[0007] Furthermore, the inputting of the rendering parameter data of the three-dimensional model of the communication room into the machine learning platform for period prediction includes: comparing the rendering parameter data of the three-dimensional model of the communication room with historical rendering parameter data; when the similarity between the rendering parameter data of the three-dimensional model of the communication room and the historical rendering parameter data is less than 50%, inputting the rendering parameter data of the three-dimensional model of the communication room into the machine learning platform for period prediction.

[0008] Furthermore, inputting the actual rendering data into the machine learning platform for learning parameter optimization includes comparing the actual rendering data with the historical rendering data, and when the similarity between the actual rendering data and the historical rendering data is less than 50%, inputting the actual rendering data into the machine learning platform for learning parameter optimization.

[0009] Furthermore, the machine learning platform in S1 predicts a rendering cycle in the future based on a time series prediction model as the optimal rendering cycle, and the time series prediction model includes a long-term and short-term neural network model or a gated recurrent unit neural network model.

[0010] Furthermore, the model building process of the optimal rendering cycle prediction algorithm based on machine learning includes: Collect historical rendering data, the historical rendering data including: the start time, end time, specific content of the rendering task, computing resources used, and network environment status of each rendering of the three-dimensional model in the communication room, Preprocessing the historical rendering data to ensure data quality and consistency; A feature set is constructed according to the historical rendering data, wherein the feature set includes directly used numerical features, constructed derivative features, resource utilization features, and external data source features.

[0011] Furthermore, the method also includes the step of training the long-term and short-term neural network model, and the step of training the long-term and short-term neural network model includes: taking the feature set as input and the optimal rendering cycle as output, optimizing the model parameters through the back propagation algorithm, and minimizing the error between the predicted value and the actual value.

[0012] Furthermore, the specific content of the rendering task includes the complexity of the model and the details of the material; the computing resources used include the CPU, GPU models and quantities; the numerical features include the number of model faces and the texture size; the derived features of the structure include time features and resource utilization features, and the time features include whether it is a weekend or a holiday; the resource utilization features include the average CPU / GPU usage rate; the external data source features include network delay and system load, and the network delay and system load can be reflected by system log data.

[0013] Furthermore, the dynamic allocation of rendering resources and dynamic selection of rendering methods for actual rendering according to the optimal rendering cycle include: adjusting the rendering resource allocation strategy in real time according to the real-time rendering requirements of the communication room three-dimensional model and the currently available resources; when it is detected that the complexity of the rendering task increases or the rendering quality requirements increase, increasing the CPU and GPU usage; when it is detected that the rendering task is relatively simple or the rendering quality requirements are not high, reducing resource allocation; when it is detected that the model to be rendered has a large number of repeated elements, using instantiation rendering technology.

[0014] Another aspect of the present application provides a communication room three-dimensional model instantiation rendering optimization system, the system includes at least one processor; and a memory storing instructions, which, when executed by at least one processor, implements the steps of any of the methods described above.

[0015] Compared with the prior art, the beneficial effects of the present invention are: Through the optimal rendering cycle of the machine learning algorithm, the rendering resources and strategies are dynamically adjusted, thus significantly improving the rendering efficiency. Through the instantiation rendering technology, high-quality rendering results can be quickly generated for specific scenes. Through the continuous learning of the machine learning model, the prediction accuracy of the rendering cycle is continuously optimized, further improving the rendering efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 : A schematic diagram of the instantiation rendering optimization method of the communication room three-dimensional model of the present application; Figure 2 : A schematic diagram of an embodiment of the communication room three-dimensional model instantiation rendering optimization method of the present application; Figure 3 : A schematic diagram of another embodiment of the communication room three-dimensional model instantiation rendering optimization method of the present application; Figure 4 : A schematic diagram of another embodiment of the communication room three-dimensional model instantiation rendering optimization method of the present application; Figure 5 : Schematic diagram of the machine learning method for instantiating and optimizing the rendering method of the three-dimensional model of the communication room of the present application; Figure 6 : Schematic diagram of the communication room three-dimensional model instantiation rendering optimization system of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] In all embodiments of the present invention, unless otherwise specified, the terms are interpreted as follows.

[0019] "Communication room" refers to a site or location where communication equipment and facilities are installed and can meet the operating conditions. Although the present invention seeks to protect the 3D model instantiation rendering optimization method of the communication room, when interpreting the protection scope of the present invention, it should not be limited to the application field of "communication room". The application of the 3D model instantiation rendering optimization method in other application fields is also within the protection scope of the present invention.

[0020] "Three-dimensional model of a communication room" refers to a model of a communication room built based on digital twin or 3D modeling technology. Unless otherwise specified, in an embodiment of the present invention, it refers to a three-dimensional model built based on an actual communication room using digital twin technology, which can perform real-time interaction and data analysis.

[0021] "Instanced rendering" refers to a rendering method (rendering algorithm) that renders repeated objects in a scene by creating multiple instances instead of rendering each object separately. The instanced rendering of the present invention is a prior art, that is, the instanced rendering algorithms in the prior art that can significantly reduce the computational burden in the rendering process are within the selection scope of the present invention.

[0022] "Optimal rendering cycle" refers to the shortest time to complete a scene rendering while ensuring the rendering quality (visual effect). It is the best matching balance point between rendering quality and performance during the rendering process.

[0023] “Rendering Resources” refers to the hardware and software resources used to complete rendering tasks, including but not limited to CPU, GPU, memory, storage space, and rendering software (rendering algorithm).

[0024] In the field of computer graphics, rendering of 3D models is an important step in generating images. Traditional rendering methods often require a lot of computing resources and time when processing complex scenes, which is particularly problematic in real-time rendering applications. For example, in scenarios such as video games, virtual reality, and augmented reality, high-quality 3D images need to be rendered quickly to provide a smooth user experience. However, existing rendering technologies are prone to performance bottlenecks when processing a large number of instantiated models, resulting in reduced rendering speed and image quality.

[0025] There are many rendering optimization methods in the prior art. For example, the patent document with application number CN202411110468.6 provides a 3D rendering method based on multi-time sequence scenes, including: collecting videos or pictures of the scene to be rendered, preprocessing the videos or pictures, using the SfM algorithm, obtaining the point cloud file and image pose information of the scene to be rendered and generating a data packet; inputting the data packet into a 3D Gaussian rendering algorithm for multi-time sequence scenes, and outputting real scene renderings and point cloud files under various viewing angles; wherein the 3D Gaussian rendering algorithm specifically includes the following steps: S1, constructing an optimized S1, Optimizer: The optimizer includes the spatial position information, spherical harmonic coefficients and transparency information of each point in the point cloud; S2, Encoding: For the existing transient objects, based on the structured information and light intensity, the point cloud and the transient encoding of the scene image to be rendered are performed; S3, Optimizer Update: All three-dimensional Gaussian points are generated according to the internal and external orientation elements of the input image to generate a simulation view from this perspective, combined with transient encoding, the transient part is removed, the loss function is calculated, and the gradient descent is used to update the optimizer; S4, Rendering: Use the various parameters of the optimizer to backpropagate the updated point cloud information, and finally obtain the rendered model file.

[0026] However, although the above rendering method improves rendering efficiency to a certain extent, it still has some limitations. For example, when processing large-scale scenes, the method is not flexible enough in handling the rendering method of objects in the scene to be rendered, and cannot adjust the rendering resources and strategies according to the characteristics of the objects, resulting in difficulty in achieving the optimal balance between rendering effect and efficiency.

[0027] Especially for the case of rendering a three-dimensional model in a communication room, due to the large number of devices and complex layout in the communication room, traditional rendering methods often find it difficult to balance rendering quality and rendering speed. In order to solve this problem, the present invention proposes a communication room three-dimensional model instantiation rendering optimization method, which can effectively balance rendering efficiency and rendering quality by optimizing the rendering algorithm and resource allocation strategy. To this end, the present application provides a communication room three-dimensional model instantiation rendering optimization method.

[0028] In some embodiments, Figure 1 and Figure 2 As shown, the present application provides, on the one hand, a method for optimizing instantiation rendering of a three-dimensional model of a communication room, the method comprising the following steps: S1: Inputting rendering parameter data of the three-dimensional model of the communication room into a machine learning platform for cycle prediction to obtain an optimal rendering cycle; the machine learning platform is constructed based on historical rendering data; The rendering parameter data in the above steps include but are not limited to the geometric detail level, texture resolution, lighting conditions, shadow effects, material properties, anti-aliasing settings, and post-processing effects of the three-dimensional model of the communication room. These parameters jointly determine the visual quality of the three-dimensional model rendering and the computational complexity of the rendering process. In the optimization method, by adjusting these parameters, the rendering cycle can be shortened and the rendering efficiency can be improved while ensuring the visual effect. For example, reducing the geometric detail level can reduce the number of polygons of the model, thereby reducing the amount of calculation during rendering; adjusting the texture resolution can balance the rendering speed and image clarity; reasonably setting the lighting and shadow effects can enhance the realism of the scene while controlling the rendering time. Through the cycle prediction of the machine learning platform, these parameters can be dynamically adjusted to achieve the best rendering effect and performance balance.

[0029] Before starting rendering, these parameters are input into the machine learning platform. The machine learning platform can simulate rendering based on these rendering parameter data to obtain the optimal rendering cycle, thereby providing personalized parameter optimization reference solutions for different rendering tasks.

[0030] S2: dynamically allocating rendering resources and dynamically selecting a rendering method to perform actual rendering according to the optimal rendering cycle, to obtain actual rendering data; the rendering method includes at least one instantiation rendering and at least one non-instantiation rendering; In this step, it can be understood that due to the complexity and changeability of the actual communication room scene, there are differences between the actual rendering data and the historical rendering data in the rendering process. In order to make the optimal rendering cycle prediction of the subsequent communication room three-dimensional model to be rendered more accurate, these actual rendering data can be fed back to the machine learning platform so that the learning parameters of the prediction model can be optimized in real time. S3: Input the actual rendering data into the machine learning platform for learning parameter optimization to obtain a machine learning platform with optimized learning parameters; the machine learning platform with optimized learning parameters performs cycle prediction based on the current rendering parameter data of the communication room three-dimensional model to obtain the next optimal rendering cycle; In the above method, learning parameters refer to the parameters used by the machine learning platform at the optimal rendering cycle, including but not limited to the model's geometric detail level, texture resolution, lighting conditions, shadow effects, material properties, anti-aliasing settings, and post-processing effects. Through learning parameter optimization, the machine learning platform can more accurately predict the rendering cycle, thereby further improving rendering efficiency and quality.

[0031] S4: Repeat steps S2 and S3 according to the next optimal rendering cycle until the actual rendering of the communication room three-dimensional model is completed.

[0032] It is understandable that the scenes in the communication room are complex and changeable. Although cycle prediction can be performed based on the machine learning platform, new variables and conditions may appear in the actual scene, which may affect the accuracy of the rendering cycle. Therefore, a dynamic adjustment mechanism is introduced in the embodiment of the present invention to adapt to the changes in rendering requirements of different scenes and ensure the accuracy of the optimal rendering cycle prediction. During the implementation process, the machine learning platform will continuously receive new actual rendering data and update the optimal rendering cycle prediction model of the machine learning platform in real time. This dynamic adjustment mechanism ensures the flexibility and adaptability of the rendering process, so that the rendering optimization method of the three-dimensional model of the communication room can be continuously improved to cope with changes in various complex communication room scenes.

[0033] Specifically, Figure 2As shown in the figure, the rendering of the three-dimensional model A in the communication room is taken as an example. According to the historical rendering data of the three-dimensional model that has been rendered in the communication room before, a machine learning platform prediction model is constructed. The machine learning platform model can predict the optimal rendering cycle required to complete the rendering of the three-dimensional model A under given rendering resource conditions after inputting the rendering parameter data of the three-dimensional model A. By analyzing the historical data, the model can identify the key factors affecting the rendering cycle and optimize the rendering process accordingly. For example, the model may find that certain specific rendering stages in the actual rendering process of the three-dimensional model A have a high dependence on the CPU, while other stages are more sensitive to the performance of the GPU. Based on these findings, the model can recommend dynamic adjustment of resource allocation during the rendering process, for example, increasing CPU resources in the CPU-intensive stage and giving priority to allocating more GPU resources in the GPU-intensive stage. In addition, the prediction model can also dynamically adjust the prediction parameters according to the actual rendering data to adapt to different rendering scenarios and hardware conditions. In this way, the rendering cycle can be significantly shortened while maintaining or even improving the rendering quality. By continuously training the model, the prediction accuracy of the model is improved, and it is constantly approaching the actual optimal rendering cycle. The rendering parameter data of the three-dimensional model A, which includes the geometric parameters, material properties, lighting conditions and viewing angle information of the object to be rendered in the communication room, the geometric parameters include the size, shape and position information of the object to be rendered; the material properties include the reflectivity, transparency and texture information of the surface of the object; the lighting conditions include the intensity, direction and color information of the light source; the viewing angle information includes the position and direction of the observer. Input into the prediction model of the machine learning platform, and the machine learning prediction model will output an optimal rendering cycle for the three-dimensional model A. In the actual rendering process thereafter, the optimal rendering cycle of A calculated by the model can be used as a reference to dynamically adjust the rendering resource allocation in the actual rendering process, and to choose whether to use the instantiation rendering technology to render the three-dimensional model A. This method of first learning the optimal rendering cycle through machine learning and then adjusting the rendering strategy according to the prediction results not only improves the rendering efficiency, but also reduces the consequences of poor rendering effects caused by improper resource allocation.

[0034] Figure 2 The data of 3D model A in the above example is obtained by inputting various rendering parameters of 3D model A into the machine learning platform before actual rendering to estimate the optimal rendering cycle of 3D model A, while the actual rendering data is collected after the actual rendering of 3D model A is completed. By comparing the estimated data of machine learning with the data after the actual rendering is completed, the model can be further learned and adjusted to improve the accuracy of future predictions of the machine learning platform.

[0035] In some embodiments, Figure 3As shown, in order to further improve the rendering efficiency of the three-dimensional model in the communication room, before the rendering parameters of the communication room three-dimensional model A are input into the machine learning platform, the rendering parameter data of the communication room three-dimensional model A is first compared with the historical rendering data. When the similarity between the rendering parameter data of the communication room three-dimensional model and the historical rendering data is less than 50%, the rendering parameter data of the communication room three-dimensional model is input into the machine learning platform.

[0036] It is understandable that it takes a certain amount of time to predict the optimal rendering cycle of three-dimensional model A using a machine learning prediction platform. Therefore, in order to improve the rendering efficiency of three-dimensional model A, the rendering parameter data of the communication room three-dimensional model A is first compared with the historical rendering data. The historical rendering parameter data of the three-dimensional model with a high similarity can be found in the historical rendering parameter data as a reference, thereby reducing the computational burden of the machine learning prediction model. In this way, an approximate optimal rendering cycle can be quickly determined for three-dimensional model A without going through the optimal cycle prediction of the machine learning platform, and the approximate optimal rendering cycle can be directly used to render three-dimensional model A, eliminating the step of going through the optimal rendering cycle of the machine learning platform, and significantly improving the rendering efficiency without having a significant impact on the rendering effect.

[0037] In some embodiments, the system presets rendering parameter templates. These templates are based on parameter settings of previous successful rendering cases. When the rendering parameter data of the communication room 3D model A is more than 50% similar to the historical rendering data, these templates can be directly applied.

[0038] In some embodiments, Figure 4 As shown, similarly, in order to avoid redundant and repeated data input into the machine learning platform, before the actual rendering data of the three-dimensional model A is fed back to the machine learning platform, the actual rendering data of the three-dimensional model A is compared and analyzed with the existing three-dimensional model historical rendering data. If it is found that these actual rendering data are repeated with the existing historical rendering data or the similarity is higher than 50%, these data can be omitted from the machine learning platform for learning parameter optimization to reduce the amount of calculation of the machine learning platform. In addition, in this way, the system can automatically identify and eliminate redundant data to ensure that the data received by the machine learning platform are the latest and valid, thereby improving the efficiency and quality of model training of the machine learning platform.

[0039] In some embodiments of the present invention, a machine learning algorithm can be used to predict the optimal rendering cycle based on the A rendering parameter data of the 3D model and external rendering resources, such as an inherent rendering engine, etc. This prediction not only takes into account the characteristics of the model itself, but also takes into account external environmental factors, such as network delay and system load, thereby minimizing the error between the predicted value and the actual value.

[0040] In some embodiments of the present application, in order to optimize the rendering process, the present application uses a machine learning-based rendering cycle prediction algorithm to learn and predict the time and resources required for future rendering by analyzing historical rendering data and multi-source data. Specifically, the present application can use a time series prediction model (such as LSTM, GRU, etc.) to predict the rendering cycle in the future.

[0041] In some embodiments of the present application, Figure 5 As shown in the figure, the influence of multiple factors such as the computer room 3D model rendering engine, multi-source data collection cycle, data processing capability, etc. is comprehensively considered, and the optimal weight coefficient is assigned to each influencing factor. The optimal rendering cycle value is obtained through calculation to provide computing power support for the real and accurate presentation of the 3D model. The specific algorithm principle is as follows: Step S1.1: Data collection and preprocessing. First, collect a large amount of historical rendering data, including but not limited to the start time and end time of each rendering, the specific content of the rendering task (such as the complexity of the model, material details, etc.), the computing resources used (CPU, GPU model and quantity), network environment conditions, etc. Then, preprocess the collected data, such as processing missing values ​​(through interpolation or deleting records with missing values), data normalization (converting data of different dimensions to the same scale), feature selection (selecting features that have a significant impact on the rendering cycle), etc., to ensure the quality and consistency of the data. This step is the basis for building the model and is directly related to the accuracy and efficiency of subsequent model training.

[0042] Step S1.2: Feature engineering. For the rendering cycle prediction problem, a suitable feature set is constructed based on historical data. In addition to the numerical features used directly (such as the number of model faces and texture size), some derivative features can also be constructed, such as time features (whether it is a weekend or holiday), resource utilization features (average CPU / GPU usage), etc. In addition, considering that the rendering process may be interfered by various external factors (such as network latency and system load), external data sources can also be introduced as features, such as system log data (reflecting the health of the system). Through feature engineering, various factors affecting the rendering cycle can be captured more comprehensively, providing richer information for the model.

[0043] Step S1.3: Model selection and training. After the features are prepared, the next step is to select a suitable machine learning model for training. For the prediction of time series data such as rendering cycles, the long short-term memory network (LSTM) is an ideal choice because it can capture long-term dependencies in the data and is suitable for processing tasks with time series characteristics. The LSTM model can effectively handle long-term dependencies and gradient vanishing problems in sequence data by introducing memory units and forget gate mechanisms. During the training process, we use the preprocessed feature data as input and the rendering cycle as output, and optimize the model parameters through the back propagation algorithm to minimize the error between the predicted value and the actual value. After the training is completed, the model needs to be verified, and the generalization ability of the model is evaluated through methods such as cross-validation to ensure that it can maintain good prediction performance on unseen data.

[0044] Step S1.4: Model optimization and deployment. After the model training is completed, a series of optimization work is often required to improve its prediction accuracy and efficiency. This includes hyperparameter tuning (such as learning rate, batch size, number of LSTM layers, etc.), regularization techniques (to prevent overfitting), ensemble learning (such as bagging, boosting), etc. Hyperparameter tuning can be performed through methods such as grid search, random search, or Bayesian optimization to find the optimal parameter combination. At the same time, considering the real-time requirements of the rendering task, the model can also be pruned, quantized, and other lightweight processing to reduce its computational complexity and memory usage, making it easier to deploy in actual environments. Finally, the optimized model is integrated into the rendering management system to automatically predict the rendering cycle and dynamically allocate computing resources.

[0045] After obtaining the prediction results of the best rendering cycle, the rendering resources and strategies are dynamically adjusted according to the prediction results, and the instanced rendering optimization technology is used to ensure a smooth interactive experience under high load. The visibility of instanced objects is controlled by modifying the number of instances. When some instanced objects need to be hidden or displayed, it can be achieved by increasing or decreasing the value of the number of instances, which greatly improves the rendering performance. At the same time, through event-driven, the availability of instanced rendering is improved, so that individual objects can be easily set to be visible and invisible, avoiding the waste of rendering resources.

[0046] In some embodiments, after the prediction result of the optimal rendering cycle is obtained, the rendering resources and strategies are dynamically adjusted according to the prediction result, and the instantiation rendering optimization technology is used to ensure a smooth interactive experience under high load.

[0047] It is worth noting that the instantiation rendering method used in this application is the instantiation rendering method provided in the patent application document with application number 202311760245.X in the prior art. The method includes constructing an instance management component, the instance management component is used to manage the subscripts of the individual object matrix in the instance matrix, and count the number of instances, setting an activation number interface for the instance management component, and the activation number interface is used to obtain the activation number of the instance; registering an activation number change event with the instance management component, updating the value of the instance number to the activation number through the activation number change event, and performing instance rendering according to the activation number.

[0048] The instantiation rendering optimization technology is particularly suitable for the communication room of the present application, because the instantiation rendering optimization technology is suitable for scenes with more repeated elements, and the communication room contains a large number of repeated elements, such as cabinets, cables, and equipment. Through the instantiation rendering technology, these repeated elements can be pre-rendered into templates, and then quickly instantiated when needed, thereby significantly reducing rendering time and improving rendering efficiency. In addition, the instantiation rendering optimization technology also supports dynamic adjustment of the number of instances to adapt to different rendering requirements and scene complexity, ensuring that a smooth interactive experience can still be provided under high load. In practical applications, this technology can effectively reduce hardware resource consumption and improve the overall performance of the rendering system, especially when processing large-scale scenes. Not only does it improve rendering efficiency, it also reduces the demand for hardware resources. By dynamically adjusting the number of instances, the system can intelligently allocate rendering tasks according to the current rendering requirements and resource conditions, thereby optimizing rendering speed and response time while ensuring image quality. In practical applications, this optimization technology can significantly improve the efficiency of the rendering process for the communication room of the present application.

[0049] For some scenes in the communication room in this application that are not suitable for instanced rendering technology, this application can also choose other existing rendering methods, such as rendering methods based on ray tracing technology. Figure 6 As shown, the present application also provides a communication room three-dimensional model instantiation rendering optimization system, the system includes at least one processor 101, a memory 102, an input device 103 and a display device 104, the input device 103 is used to obtain input from the outside world, the memory 102 stores instructions, and when the instructions are executed by at least one processor 101, the steps of the method described in the method embodiment are implemented, and the running results are displayed on the display device 104, and the steps of the method described in the method embodiment are implemented.

[0050] In one embodiment of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the operating activity method for issuing customized rewards to users based on different trigger conditions described in any of the above embodiments are executed.

[0051] The embodiments of this document also provide a computer-readable instruction, wherein when a processor executes the instruction, the program therein causes the processor to execute the steps of the operation activity method of issuing customized rewards to users based on different trigger conditions as described in any of the above embodiments.

[0052] Although this specification contains many specific implementation details, these should not be interpreted as limitations on the scope of any invention or the scope of what can be claimed, but rather as illustrations of features that can concretize specific embodiments of specific inventions. Specific features described in this specification in the context of independent embodiments can also be implemented in combination with a single embodiment. Conversely, the various features described in the context of a single embodiment can also be implemented independently in multiple embodiments, or in any suitable sub-combination. In addition, although the features can be described above as a combination effect and even initially required, one or more features from the required combination can be removed from the combination in some cases, and the required combination can be turned to a sub-combination or a variation of the sub-combination.

[0053] Similarly, although operations are described in the accompanying drawings in a particular order, it should not be understood that: in order to achieve the desired results, such operations are required to be performed in the particular order shown or in a sequential order, or that all illustrated operations are performed. In certain cases, multitasking and parallel processing can be advantageous. In addition, the separation of various system modules and components in the above-mentioned embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that program components and systems can generally be integrated into a single software product or packaged into multiple software products.

[0054] Specific implementations of the subject matter have been described. Other implementations are within the scope of the following claims. For example, the activities recited in the claims can be performed in a different order and still achieve the desired results. As an example, the processes described in the accompanying drawings do not necessarily require the specific order or sequential order shown in order to achieve the desired results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

1. A method for optimizing instantiation rendering of a three-dimensional model of a communication room, characterized in that: The method comprises: S1. Inputting the rendering parameter data of the three-dimensional model of the communication room into a machine learning platform for cycle prediction to obtain an optimal rendering cycle; the machine learning platform is constructed based on historical rendering data, and the historical rendering data includes historical rendering parameter data; S2. Dynamically allocate rendering resources and dynamically select a rendering method to perform actual rendering according to the optimal rendering cycle to obtain actual rendering data; the rendering method includes at least one instantiated rendering and at least one non-instantiated rendering; S3, inputting the actual rendering data into the machine learning platform to optimize the learning parameters, and obtaining a machine learning platform with optimized learning parameters; the machine learning platform with optimized learning parameters performs cycle prediction based on the current rendering parameter data of the three-dimensional model of the communication room, and obtains the next optimal rendering cycle; S4. Repeat steps S2 and S3 according to the next optimal rendering cycle until the actual rendering of the communication room three-dimensional model is completed.

2. The communication room three-dimensional model instantiation rendering optimization method according to claim 1, characterized in that: The rendering parameter data of the three-dimensional model of the communication room includes: geometric parameters, material properties, lighting conditions and viewing angle information of the objects to be rendered in the communication room, the geometric parameters include the size, shape and position information of the objects to be rendered; the material properties include the reflectivity, transparency and texture information of the object surface; the lighting conditions include the intensity, direction and color information of the light source; the viewing angle information includes the position and direction of the observer.

3. The communication room three-dimensional model instantiation rendering optimization method according to claim 1, characterized in that: The S1 also includes: comparing the rendering parameter data of the three-dimensional model of the communication room with the historical rendering parameter data, and when the similarity between the rendering parameter data of the three-dimensional model of the communication room and the historical rendering parameter data is less than 50%, inputting the rendering parameter data of the three-dimensional model of the communication room into the machine learning platform for period prediction.

4. The communication room three-dimensional model instantiation rendering optimization method according to claim 1, characterized in that: S3 specifically includes comparing the actual rendering data with the historical rendering data, and when the similarity between the actual rendering data and the historical rendering data is less than 50%, inputting the actual rendering data into the machine learning platform for learning parameter optimization.

5. The communication room three-dimensional model instantiation rendering optimization method according to claim 1, characterized in that: The machine learning platform in S1 predicts a rendering cycle in the future based on a time series prediction model as the optimal rendering cycle. The time series prediction model includes a long-term and short-term neural network model or a gated recurrent unit neural network model.

6. The communication room three-dimensional model instantiation rendering optimization method according to claim 5, characterized in that: The method further includes the step of establishing the time series prediction model, wherein the step of establishing the time series prediction model includes: Collect historical rendering data, the historical rendering data including: the start time, end time, specific content of the rendering task, computing resources used, and network environment status of each rendering of the three-dimensional model in the communication room, Preprocessing the historical rendering data to ensure data quality and consistency; A feature set is constructed according to the historical rendering data, wherein the feature set includes directly used numerical features, constructed derivative features, resource utilization features, and external data source features.

7. The communication room three-dimensional model instantiation rendering optimization method according to claim 6, characterized in that: The method also includes the step of training the long-term and short-term neural network model, wherein the step of training the long-term and short-term neural network model includes: taking the feature set as input and the optimal rendering cycle as output, optimizing the model parameters through a back propagation algorithm, and minimizing the error between the predicted value and the actual value.

8. The communication room three-dimensional model instantiation rendering optimization method according to claim 6, characterized in that: The specific content of the rendering task includes the complexity of the model and the material details; the computing resources used include the CPU, GPU models and quantities; the numerical features include the number of model faces and the texture size; the derived features of the structure include time features and resource utilization features, and the time features include whether it is a weekend or a holiday; the resource utilization features include the average CPU / GPU usage rate; the external data source features include network latency and system load, and the network latency and system load can be reflected by system log data.

9. The communication room three-dimensional model instantiation rendering optimization method according to claim 1, characterized in that: The dynamic allocation of rendering resources and dynamic selection of rendering methods for actual rendering according to the optimal rendering cycle include: when it is detected that the complexity of the rendering task increases or the rendering quality requirement increases, increasing the usage of the CPU and GPU; when it is detected that the rendering task is relatively simple or the rendering quality requirement is not high, reducing the allocation of rendering resources; when it is detected that there are many repeated elements in the model to be rendered, using instantiation rendering technology.

10. A communication room three-dimensional model instantiation rendering optimization system, characterized in that: The system comprises at least one processor; and a memory storing instructions, which, when executed by the at least one processor, implement the steps of the method according to any one of claims 1 to 9.

Images